The present invention relates generally to computer software and, more particularly, to a client management product for a computer system.
Generally, it is desirable to utilize client management products, such as capacity planning products, to proactively monitor one or more client machines, especially in large enterprise configurations utilizing multiple machines dedicated to specific functions within the network. For example, a capacity planning product can determine when an existing server system will reach or exceed the limits of its hardware. In comparison, a sizing product, unlike a capacity planning product, involves a determination of the hardware investment needed to create a new computer system. Other client management products (namely, performance monitoring products) are used to monitor the performance of computer networks and systems. Such performance monitoring products provide real-time monitoring of performance thresholds.
Traditionally, capacity planning products and performance monitoring products are implemented separately. However, running simultaneously, those products can generate significant levels of administrative traffic on a network. Additionally, the system administrator must monitor two distinct client data gathering programs. This is especially inefficient if both are utilizing the same data from the client to implement their function. Accordingly, there is a need for a product which integrates the functions of capacity planning and performance monitoring into a single client management product which more efficiently utilizes system resources.
Existing client management products sometimes utilize polling agents to collect data from a computer network and its components. One approach is to provide a central polling agent which issues data requests for client data (e.g., capacity planning or performance monitoring data) from a number of network machines at given intervals. An alternative approach is to provide a number of resident polling agents which reside on the network machines and which are responsible for transmission of client data to a central database at similar given intervals. Some network administrators prefer the central polling agent configuration because it does not require the system administrator to load and maintain polling software on each client. Other network administrators prefer to use resident polling agents because it reduces the amount of network traffic relating to the collection of capacity planning data. Currently, a network administrator typically implements one polling agent configuration without the ability to change the configuration as management needs vary. Given the demand for both approaches, there is a need for a client management product which allows a user such as a network administrator to select from multiple polling functions or modes, including a central polling function and a resident polling function.
Client management data, such as capacity planning data, is obtained for a given client by taking a number of measurements involving various system components. Some operating systems provide functions which measure and transmit various system parameters. For example, several operating system vendors support the Web-Based Enterprise Management (WBEM) standardized format for accessing and sharing management information. Counters, which are stored values that are incremented every time an event occurs, or after a measurement interval has been obtained, are often used to measure memory usage or CPU (central processing unit) usage on individual machines and to monitor utilization or xe2x80x9cbusy timexe2x80x9d for links. Historical information from counters can be used for a number of purposes, including the identification of bottlenecks in performance (such as extensive disk queue lengths), the detection of periods of peak usage and the prediction of future usage. A measurement typically involves a grouping or selection of multiple counters or other such devices.
Conventionally, an active measurement is initiated by a monitoring process that identifies which measurements should be produced by the client. However, once initiated, the measurement cannot be modified except by deactivating it, making the modification and then reactivating it. This causes a discontinuity in the collection of client management data and can result in the loss of valuable information while the measurement remains inactive. Moreover, substantial user intervention may be required to complete the modification. Consequently, there is a need for a client management product which provides the ability to modify an active measurement without deactivating it.
Because separate client management products have been used to collect capacity planning data and performance monitoring data, the data obtained from active measurements is typically forwarded to separate databases at a central location without first being stored in the client""s cache or other local memory. Capacity planning data is usually not cached because it is collected infrequently and would consume the client""s cache or be overwritten by other data. Performance monitoring data is typically not cached because it is critical data which should be monitored frequently. However, if the performance data were stored locally and averaged over a period of time, it would provide more accurate capacity planning data because it would be based on an increased number of data points. Accordingly, there is a need for a client management product which stores performance data locally and averages it over a period of time so that more accurate capacity planning data is provided to the central location.
Many service transactions are under a xe2x80x9cservice level agreementxe2x80x9d (SLA), which is a contract between the provider of computer or database services and user of those services. The SLA specifies the minimal level of service that is guaranteed and usually is in the form of maximum response times for certain transaction. Violations of an SLA is likely to occur if no measures are taken to anticipate potential violations. Accordingly, a capacity planning product can be utilized to anticipate system needs with respect to SLAs.
In addition to SLAs, service transactions may be subject to other time constraints, and it would be advantageous to monitor the response time in a service chain, for example, to determine in advance whether the time constraints will likely be met. A xe2x80x9cservice chainxe2x80x9d is a collection of resources that are used in the processing of a transaction or service. Because of the real-time need for monitoring, a separate performance monitoring product would also need to be utilized. The substantial deficiencies associated with concurrently running two client management products evidences a need for a client management product which can monitor multiple aspects of individual service transactions in a computer system by identifying and monitoring the specific system resources involved.
It is oftentimes desirable to archive capacity planning and performance monitoring data. Although archiving requires substantial memory, substantial resources were expended to collect the data in the first place and the archived data may be valuable for predictive analysis and for generating reports based on detailed historical data to justify an expensive upgrade. Traditionally, the archiving of capacity planning data cannot be carried out within the client management tool itself. Instead, a conventional method of archiving must be used, which must be carried out utilizing software applications outside of the client management tool and which typically requires the expertise of a database specialist. Thus, there is a need for a client management product which performs both a client management function, such as capacity planning, and an archiving function without the need for a database specialist.
Generally, conventional management tools (e.g., capacity planning or performance monitoring tools) store data collected from one or more clients into database files. However, depending on the management tool selected and the provider of the management tool, the data can be organized under a variety of database column headings, which are often proprietary in nature. For example, a first database file may have a column heading entitled xe2x80x9cMachine Namexe2x80x9d to represent the specific identity of a server while a second database file may utilize the column heading xe2x80x9cComputer Namexe2x80x9d to represent the same information. Accordingly, data collected via different management tools and/or different tool providers may be incompatible because of inconsistencies in the data schema. Thus, there is a need for a management tool that can read and/or convert previously collected management data.
Additionally, because none of the conventional management tools are used for both capacity planning and performance monitoring, the collected data is saved in variable length files, known as a log file, whose length cannot be adjusted, or added to, once the measurement is closed. For example, Microsoft Corporation provides a PERFMON performance monitoring function in its WINDOWS NT brand operating system. Typically, PERFMON can save the collected data into an EXCEL spreadsheet format, which is unique to Microsoft products. However, once a PERFMON measurement is stopped, a reactivation of the same measurement criteria instigates the creation of a second log file having the same name. Currently, there is no method, absent saving the contents of the two separate files to a third file, for storing additional management data on the original file. Thus, there is a need for a management tool that can store multiple measurement data within the same measurement file.
Another drawback to conventional capacity planners is their user interface (UI), which typically requires a highly trained system administrator or system manager. For example, many conventional UIs do not allow a user to manipulate the system in a graphical form, relying on generally more difficult command-based interfaces. Moreover, the UIs generally entail multiple screen interfaces to carry out a single function of the management system. These aspects of a conventional UI require more system administrator training and thus increase the costs associated with a capacity planner. Accordingly, there is also a need for a capacity planning product having a more intuitive UI which reduces the costs associated with the product because less training is required for the system administrator.
Generally described, the present invention provides a method in a computer system for implementing one or more polling agents for collecting data from a plurality of computer system objects. According to the method, a user is prompted to select one of a plurality of polling functions. If a user selection is received, at least one polling agent is implemented corresponding to the selected polling function.
In another aspect of the invention, a computer system for managing a plurality of objects is provided. The system includes a central manager and a first of the plurality of objects being in communication with the central manager. Additionally, the system includes a polling agent for collecting data from the plurality of objects. The polling agent resides on the first object when the computer system operates in a first polling mode and the polling agent resides on the central manager when the computer system operates in a second polling mode.
In a further aspect of the invention, a computer system for managing a plurality of objects is provided. The system includes means for centrally managing the computer system and a first of the plurality of objects being in communication with the central managing means. Additionally, the system includes means for collecting data from the plurality of objects. The collecting means reside on the first object when the computer system operates in a first polling mode and the collecting means reside on the central manager when the computer system operates in a second polling mode.
A method and system is provided for implementing a variety of polling agent functions in a client management tool. Users at a central manager are prompted to select either a central polling function or a resident polling function for the collection of client management data from a number of computer objects. Additionally, the polling agent functions of the present invention are capable of distinguishing between critical data and non-critical data and determining an appropriate transmission interval for the data.